Smart cushion: A practical system for fine-grained sitting posture recognition

Abstract

Poor sitting postures influence one's health and can cause upper limb and neck disorder. Current solutions for siting posture recognition, however, are impractical due to intrusiveness, high cost or low generalization capability. Particularly, most of the existing solutions are chair-dependent, which are highly coupled with certain types of chairs. In this paper, we design Postureware, a smart cushion, which is a low-cost, non-intrusive and general sitting posture recognition system. In particular, Postureware incorporates very thin pressure sensors to offer non-intrusive experience, an effective sensor placement solution to reduce cost, a set of user-invariant features and an ensemble learning classifier to improve generalization ability. We implement a prototype system and conduct extensive experiments. The results show that Postureware can classify fifteen fine-grained postures with high accuracy.